Profiling blood donors in Egypt: A neural network analysis Mohamed M. Mostafa Auburn University, College of Business, 242 Lowder Business Building, 415 West Magnolia Avenue, Auburn, AL 36849, USA article info Keywords: Blood donation Consumer profiling Neural networks Discriminant analysis Egypt abstract This study uses intelligent modeling techniques to examine the effect of various demographic, cognitive and psychographic factors on blood donation in Egypt. Two artificial neural network models (multi-layer perceptron neural network [MLP] and probabilistic neural network [PNN]) are compared to a standard statistical method (linear discriminant analysis [LDA]). The variable sets considered are sex, age, educa- tional level, altruistic values, perceived risks of blood donation, blood donation knowledge, attitudes toward blood donation, and intention to donate blood. The paper shows how it is possible to identify var- ious dimensions of blood donation behavior by uncovering patterns in the dataset, and also shows the classification abilities of two neural network techniques. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction Blood transfusion is a universally practiced, remarkably safe, routine clinical procedure. Blood is needed not only in emergency situations but also for routine operations, such as heart bypasses, and medical treatments. Leukaemia patients, for example, can use up to 700 donations over the course of their treatment (Barkworth, Hibbert, Horne, & Tagg, 2002). A well-functioning blood transfusion service is dependent on forthcoming blood do- nors who are willing to donate voluntarily (Sandborg, 2000). The recruitment of blood donors is, however, a challenge. For example, though 40,000 Americans donate blood each day, that is barely enough to keep the health care system running. A major reason for the blood shortage is that though 60% of the US popula- tion is eligible to donate blood, only about 5% does (National Woman’s Health Report, 2003). In developing countries traditional culture beliefs as well as the lack of knowledge and technical expertise can often frustrate efforts to achieve a truly voluntary blood donor system (Zaller et al., 2005). For example, in Egypt the National Blood Transfusion Center (NBTC) receives an average of just 60 donors every day. There are currently some 250 blood banks in government hospitals and 13 in private hospitals (Leila, 2007). Blood donation rates in Egypt are, however, low by interna- tional standards as only 1.5% of the population donates blood (El-Jesri, 2007). Although there are several studies on attitudinal influences of blood donors and non-donors in developed countries (e.g. Andaleeb & Basu, 1995; Royse & Doochin, 1995; Sojka & Sojka, 2002) and in developing countries (e.g. Hosain, Anisuzzama, & Begum, 1997; Wiwanikit, 2000; Zaller et al., 2005), there are no published data concerning the most important factors relating to blood donation in Egypt. In this research we fill this research gap by profiling blood donors and non-donors in Egypt based on demo- graphic characteristics, personality traits, attitudes and various types of motivation. This paper will first present the research objectives followed by relevant literature review and hypotheses development. Next, methodology and results will be discussed. Implications section follows. Finally, research limitations and directions for future re- search will be explored. 2. Research objectives Profiling blood donors is very important because the first step in planning the target marketing strategy is to segment the market and develop profiles of the resulting market segments. In fact, the usefulness of market segmentation hinges upon accurate pro- filing. Relatively low accuracy in forecasting segment membership will result in ineffective marketing programs and potential nega- tive impact due to targeting unintended segment members. De- spite the importance of profiling blood donors, researchers have largely ignored blood donation specific topics related to consumer behavior, values, and culture. From a public policy standpoint, it is important to know what motivates individuals to donate blood if a pro-blood donation change policy is to be successfully imple- mented. Using intelligent modeling techniques, this research aims at profiling blood donors versus non-donors. More specifically, the aim of this study is twofold: (1) to investigate the influence of var- ious demographic, cognitive and attitudinal factors on blood dona- tion behavior in Egypt; and (2) to compare the classification performance of NNs against the more traditional techniques such as LDA within the context of blood donation behavior. 0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.eswa.2008.06.048 E-mail address: mmm0010@auburn.edu Expert Systems with Applications 36 (2009) 5031–5038 Contents lists available at ScienceDirect Expert Systems with Applications journal homepage: www.elsevier.com/locate/eswa